The supporting processes outlined below are considered to be important or noteworthy in the context of maintaining the character of the site, but are not considered to represent critical processes in the context of the considerations outlined in section 3.1.1 of this report. In this context:
Supporting processes may operate over broad spatial scales and are not considered likely to be fundamentally altered by activities within the site.
Some supporting processes are already partially covered by other critical components, processes or services/benefits.
The supporting processes, while not critical, are important to wetland functioning and are noteworthy in this regard.
Climate conditions for the Alligator Rivers Region, have been described by Saynor et al. (2000 and references therein). In general, the climate of the Alligator Rivers Region can be defined as wet-dry tropical with a wet season duration of four-and-a-half to seven months. Humidity is generally highest between January and March with mean relative humidity (at 9 am) greater than 80 percent. Temperatures at Jabiru are high throughout the year (Figure 3 -35), with temperatures higher than 30 degrees Celsius observed on average more than 320 days per year between 1971 and 2009. Annual mean minimum and maximum temperatures were 22.5 degrees Celsius and 34.2 degrees Celsius, respectively. The highest temperatures are generally recorded from September to October while lowest temperatures usually occur from June to August (Figure 3 -35). Average annual maximum temperatures in the Northern Territory have increased by about 0.12 degrees Celsius per decade since 1950 together with an increase in frequency of extremely warm days and nights and a concurrent decrease of extremely cool days and nights (Hennessy et al. 2004). Greater warming was observed in May to October compared to November to April.
The warmer wet season is marked by monsoonal depressions bringing heavy rain and occasional tropical cyclones to the area. Over 90 percent of the average rainfall occurs during the wet season between November and March (Figure 3 -36) with mean annual rainfall ranging from approximately 1300 millimetres in the south to 1500 millimetres at Jabiru. Little or no rain occurs during the cooler dry season extending from May to September (Figure 3 -36). Potential evaporation (2400 – 2700 millimetres per year) exceeds rainfall in most years (Saynor et al. 2000).
The Northern Territory has become wetter between 1950 and 2002 with average rainfall rising 35.7 millimetres per decade for November to April and falling 0.4 millimetres per decade for April-October (Hennessy et al. 2004). Particularly strong rainfall periods were observed in the mid-1970’s and in 1999-2000 and the highest rainfall on record at Jabiru Airport was observed during the wet season of 2007 (Hennessy et al. 2004, Bureau of Meteorology website 2009).
Winds are predominantly from the south-east and east between April and September, whereas winds are more variable with an often strong westerly and northerly component from November to February.
Figure 3 35 Mean maximum and minimum temperature at Jabiru Airport between 1971 and 2009. Upper and lower error bars denote the 90th and 10th percentiles (source: Bureau of Meteorology unpublished data)
Figure 3 36 Mean monthly rainfall and mean number of rain days at Jabiru Airport between 1971 and 2009. Upper and lower error bars denote the 90th and 10th percentiles (source: Bureau of Meteorology unpublished data)
The El Niño Southern Oscillation (ENSO) modulates the behaviour of the monsoon and frequency of cyclones experienced (Hennessy et al. 2004, Wasson et al. in prep.). The El Niño phase tends to suppress monsoon and cyclone activity over the Northern Territory, while the La Niña phase tends to enhance this activity. Hence, dry periods tend to be El Niño years, whereas the wet periods are usually La Niña years.
However, further climate variability on longer, decadal time scales was suggested by Power et al. (1999). In particular, the Inter-decadal Pacific Oscillation (IPO) has been shown to be associated with decadal climate variability over parts of the Pacific Basin, and to modulate interannual ENSO-related climate variability over Australia (Salinger et al. 2001). Accordingly, Wasson et al. (in prep.) noted an approximate 20-year decadal variation in rainfall and flow for the Magela Creek (18 years) and Katherine River (22 years) catchments. The decadal variations in rainfall and flow were essentially in-phase. Bayliss et al. (2008) also noted that other major rivers across the “Top End” of the Northern Territory exhibit 20-25 year periodicities in flow volume.
Such long term decadal periodicities in rainfall and flow may have important implications for the biology in the area. This was demonstrated by Bayliss et al. (2006, 2008) who showed that magpie geese Anseranas semipalmata across the Northern Territory exhibited approximately 20 year population cycles that were coupled to similar and generally coherent periodicities in flow of the Katherine River, Daly River and Magela Creek (refer Figure 3 -37). The authors noted an average response time lag of three to five years between river flow and magpie goose numbers.
Figure 3 37 Cusum plots (cumulative sum of mean deviations) of magpie goose numbers (white symbols) in the Northern Territory and Katherine River flow (black symbols) Figure reproduced from Bayliss et al. (2008)
188.8.131.52Extreme Climatic Events
As would be expected for a tropical locality, the Ramsar site experiences severe weather events in the form of cyclones and the associated strong winds, heavy rains and destruction. An average of eight to ten cyclones form annually between November and May in tropical northern Australia (Holland 1984). Monthly totals of tropical cyclones in the Northern Territory over the last four decades are shown in Figure 3 -38.
A recent example of an extreme climatic event is a tornado that swept through the National Park in March 2007, with winds of between 230 and 270 kilometres per hour that left a three kilometre path of snapped or uprooted trees and damaged caravans near the Mary River Ranger Station, and record-breaking rainfall that flooded the Oenpelli and Adelaide River areas3.
It is highly likely that the frequency and severity of extreme climatic events will increase as a result of climate change (see Section 5.3 and BMT WBM 2010).
Figure 3 38 Total number of tropical cyclones in the Northern Territory between 1963 and 2006 by category, where C5 is the most destructive category (source: Bureau of Meteorology)
3.6.2Geology and Geomorphology
Landscape features within Kakadu National Park cover over 2000 million years of geological evolutionary history (Press et al. 1995b). The Kakadu National Park region is situated in the eastern part of a major geological structure known as the Pine Creek Geosyncline, which is the main geological structure of the region (Commonwealth of Australia 1988, Press et al. 1995b). The region provides a favourable setting for mineral deposits of economic significance, particularly in the eastern part.
In the historic Stage I and Stage II areas of Kakadu National Park, the mineralisation is mainly located in a rock formation known as the Cahill Formation, which includes the Ranger, Jabiluka and Koongarra uranium deposits. The Cahill Formation is considered a major uranium deposit on a world scale, but exploration indicated the presence of a range of metals, including gold (Commonwealth of Australia 1988). Mineralisation in the Stage III area of the region is primarily located in a series of rock formations known as the El Sherana Group and includes gold, platinum and minor uranium deposits.
While the area is regarded as highly prospective for mineral exploration, the rock formations are often masked by younger overlying sequences, including the sandstones forming the Arnhem Land escarpment. The sandstone formations of the Arnhem Land escarpment date back about 1700 million years and can be up to 300 metres in depth. Erosion of the younger rocks has resulted in a deep soil cover over much of the region, which is about 40 metres deep. However, the tectonic stability and antiquity of the landscape resulted in deep weathering of the rocks forming strongly leached and relatively infertile soils (Press et al. 1995b). Together with the pronounced rainfall seasonality, this has markedly influenced flora and fauna development in the region.
In contrast, the extensive coastal and riverine alluvial plains are of recent origin, often dating back no more than a few thousand years (Press et al. 1995b). These recent landforms are a result of sediment deposition in drowned river valleys associated with sea level stabilisation at about its present level ca. 6000 years ago. Hence, the deeper saline sediments underlying the floodplains are overlain by brackish, organic-rich, acidic soils, which support the freshwater wetlands (Press et al. 1995b).
Rivers in the Northern Territory have several morphological phases and move from one phase to another as they respond to tidal pressure and seasonal freshwater runoff (Chappell and Woodroffe 1984, Petty et al. 2005). Different longitudinal regions of the rivers will exhibit distinct morphological features depending on the state of development of the particular region (Figure 3 -39).
Because the flooding tide has higher peak velocities than the ebbing tide, a much higher sediment load can be transported during floods. This sediment is deposited along the tidal channels, gradually forming mud levees at the upper estuary (refer Figure 3 -39). These levees contain the channel, prevent further saltwater penetration and impound freshwater in large wetlands (Petty et al. 2005). During the course of the dry season, the salinity of these wetlands will increase, and in some areas will become quite brackish.
The annual freshwater impoundment maintains the low salinity soil surface of the wetlands, overlying a highly saline subsoil region. Without the impoundment, subsoil salt may emerge resulting in widespread die-off of freshwater species.
Some notable morphological features formed in recent times (several thousand years) include palaeo-channels, dendritic channels and billabongs, which generally form in palaeo-channels. Palaeo-channels are remnant tidal channels that were active during the mid-Holocene, and have since been partially or completely infilled by deposition of tidal sediments (Woodroffe and Mulrennan 1993). They are apparent as billabongs, freshwater swamps and wetlands. As palaeo-channels are some of the lowest-lying topography within a coastal floodplain, they are particularly vulnerable to saltwater intrusion. The intrusion of saltwater can result in the death of freshwater vegetation and development of bare surfaces susceptible to aeolian (wind blown) erosion.
Figure 3 39 Typical cross section of (top) upper estuary and (bottom) lower estuary
Tidal processes of rivers in the region have been described in detail by Vertessy (1990) and are summarised here. In the South Alligator River, the largest river within the site, the tidal component extends from the mouth in Van Diemen Gulf to just downstream of Yellow Water for a distance of about 105 kilometres. The maximum range in tide height at the mouth is 5.8 metres. Comparable values at other stations along the South Alligator River indicate that there is only minor attenuation of tidal amplitude with distance from the sea over most of the river.
Tides are moderately asymmetric at the mouth (that is, not equal in duration), with spring tide ebb and flood durations of approximately 415 and 320 minutes and neap tide ebb and flood durations of 410 and 325 minutes, respectively. Ebb/flood duration and velocity become increasingly asymmetric with distance from the mouth, with flood tides having shorter durations and higher peak velocities. In the mid and upstream reaches flood current velocities approach two metres per second, whereas ebb current velocities rarely exceed one metre per second. Due to the shorter flood tide with a steeper gradient, the flood tide can carry a much higher sediment load than the ebb tide, consequently resulting in sediment transported up the river.
Storm surges induced by cyclonic winds in the order of four metres are another major influence on water levels at the mouth of the South Alligator River. This increase in water level is added to the existing tide level at the time of the cyclone to give a combined storm tide level. Because of the random nature of the combination of tide and storm surge, however, this level is not necessarily above the Highest Astronomical Tide level (HAT) for the river. With regards to salinity, flows experienced during the wet season are sufficient to flush the tidal channel to fresh water levels over almost the full length of the estuary (see BMT WBM 2010).
Description and Patterns in Variability
Only limited water quality data are available from the Alligator Rivers Region. While water quality data are available from a number of stations within the Magela Creek catchment in the eastern part of Kakadu National Park, only little information on water quality is available from the western and middle sections of the site (Figure 3-19). Most of the data were collected between the early 1970s and mid-1980s. Since this time, there has been some intensive monitoring in selected waterways in the Magela floodplain; however there has been comparatively little water quality data collection elsewhere in the site.
There are no regional or local water quality guidelines for the Ramsar site. Therefore, available water quality data have been compared with ANZECC/ARMCANZ (2000) guidelines values. However, it is to be noted that the ANZECC guidelines are not based on regional reference values and consequently any non-conformity with the ANZECC guidelines does not necessarily represent a change in ecological character (refer Section 4.3).
Table 3-7 shows a comparison of the 20th and 80th percentiles of available water quality data with ANZECC/ARMCANZ (2000) guideline values. It is noted from this data that the water in streams of the region are fairly acidic, with the 20th percentile of pH often below the ANZECC guideline value. Relatively high acidity is especially notable for stations along the Magela Creek Plain, characterised by acid sulfate soils and where extremely acidic “first flushes” after the dry season are commonly observed events (Willett 2008).
While no information on nutrient concentrations is available for the upland stations of the region, the 90th percentile of nutrient concentrations often exceeded ANZECC guideline values at the lowland stations (Table 3-7). Notably high 80th percentile nutrient concentrations were recorded at stations within the Magela Creek catchment, a region characterised by a number of backflow, channel and flood plain billabongs (Hart and McGregor 1980). Concurrent with increased nutrient concentrations, elevated 80th percentiles of conductivity, turbidity and chlorophyll a and relatively low 20th percentiles of dissolved oxygen were recorded at these stations (refer Table 3-7). In line with these findings, Hart and McGregor (1980) showed that billabong waters in the Magela Creek catchment had elevated conductivity, turbidity, nutrient and chlorophyll a concentrations at the end of the dry season. A significant improvement in water quality with commencement of the wet season and flushing of the billabongs was noted. The authors also demonstrated that billabong waters became significantly depleted in dissolved oxygen when mixing of the water column did not occur for several days under certain conditions.
In support of these results, Figure 3-20 shows the seasonality in water quality parameters at a site within the Magela Creek catchment. Conductivity is generally higher around the end of the dry season, concurrent with an increase in nutrient concentrations. Based on these results, evapo-concentration is the likely mechanism that leads to the observed increase in nutrient and ion concentrations in the Magela Creek catchment. As a result of increased nutrient supply, chlorophyll concentrations may increase within the billabongs. With the onset of wet season flushing, nutrient concentrations are diluted as demonstrated by the concurrent decrease in conductivity values (Figure 3-20).
Townsend and Douglas (2000) investigated the effect of different fire regimes on stream water quality in Kakadu National Park in three Eucalypt-dominated open-forest catchments. The authors found that fires lit in the late dry season reduced canopy cover, riparian tree density and increased the amounts of bare ground, thereby increasing erosion. Accordingly, Townsend and Douglas (2000) found that storm runoff concentrations of suspended sediments, iron and manganese were two to five times higher after late dry season fires. In contrast, fires lit during the early dry season had an overall negligible effect on wet season stream water quality, except possibly for a small increase in nitrogen concentrations and loads (Townsend and Douglas 2004).
Estuarine and Marine Waters
There are no available empirical data describing the water quality of estuarine and marine waters. This is considered an important information gap.
Table 3 16 Water quality data (80th percentiles; 20th and 80th percentiles for dissolved oxygen and pH) from gauging stations in the Kakadu National Park catchment and comparison to ANZECC guideline values. Values in red denote exceedance of guideline limits. (source: NRETAS unpublished data)
Lowland stations: less than150 metres; upland stations: greater than 150 metres or directly adjacent to 150 metres contour (refer Figure 3-19) Note that % dissolved oxygen saturation values were calculated according to Weiss (1970) from measured concentrations in mg/L and the median temperature at the respective sites (salinity = freshwater)
I:\B17399_I_GML Kakadu Ramsar GWF\DRG\ECO_004_091116 Kakadu WQ Sites.wor
Figure 3 40 Water quality gauging stations in Kakadu National Park
Figure 3 41 Seasonal pattern of conductivity and nutrient concentrations (total nitrogen and ammonia) at gauging station G8210017
NRETA has produced an overall map of the groundwater resources of the Northern Territory (refer Ticknell 2008). Aquifer types within Kakadu National Park have been classified as the following as shown in Figure 3 -42:
Fractures and Weathered Rocks with local scale aquifers which occur broadly across the site in the floodplain and catchment areas.
Fractures and Weathered Rocks with minor groundwater resources and local scale aquifers which occur over the majority of the escarpment Stone Country.
Fractured and Karstic Rocks with intermediate to local scale aquifers which occur over small areas in the north-eastern and southern parts of the escarpment.
Sedimentary Rocks with intergranular porosity with regional to local scale aquifers which occur along the lower reaches of the East Alligator River and within an area of the south-west of the site.
The scale of aquifers as mentioned above refers to distance over which groundwater flows through the aquifer from recharge to discharge areas. Specifically, local scale indicates less than five kilometres, intermediate scale refers to distances of five to 50 kilometres, and regional scale refers to greater than 50 kilometres.
As outlined previously in the discussion of wetland types, Kakadu National Park has several groundwater-dependant freshwater springs that are predominantly situated in the northern-eastern portion of the Stone Country between the South and East Alligator Rivers. A number of seeps have also been identified by NRETA further south in the Stone Country but no empirical data on their exact location are available. Such features would likely correspond with the larger scale aquifers present in these environments.
Groundwater usually discharges at low lying points in the landscape. It can take the form of individual springs or as diffuse seepage into streams. Dry season flows into the river systems of the Northern Territory are maintained through groundwater discharge particularly in karstic systems where there are more substantial groundwater resources present. However, as these karst systems are very limited within the Ramsar site, it is likely that the contribution of groundwater to surface water flows is minor for the mid and lower reaches of Kakadu’s major river systems and/or is highly localised in relation to major billabongs. Finlayson et al. (2006) indicates that in the floodplain areas, groundwater levels can fall between two to four metres during the Dry Season.
Groundwater discharge may play a much more important role in the context of the upper reaches of the river systems and for permanent and ephemeral streams located in the upstream and escarpment areas but no studies to date have assessed the interaction between groundwater and surface water in these environments and the relative contribution of groundwater flows during the dry season.
Groundwater discharge can also occur through trees and riparian vegetation tapping directly into shallow water aquifers. This process is identified as being very significant in the southern parts of the Northern Territory by Ticknell (2008), but is not identified as a dominant process in the northern, coastal areas of the Territory such as Kakadu. The extent to which this process occurs in the site is also an information gap.
While there have not been any comprehensive studies of groundwater for the Park to date, there have been site-specific studies of the mineral lease area in the context of movement of potential contaminants from Jabiluka mine tailings by groundwater flow toward the park (refer Kalf and Dudgeon 1999). In this study groundwater flows from the mine site into the Magela floodplain was modelled to predict the concentrations of contaminants to be expected along the flow paths. In terms of groundwater movement, the findings of the study noted that there were weak upward components of groundwater flow both east and west of the mine and it was considered that any such flow which reaches the shallow alluvial or weathered rock zone would be diluted and flushed away by the annual wet season surface flows.
I:\B17399_I_GML Kakadu Ramsar GWF\DRG\ECO_008_100223_ Groundwater_Aquifers.wor
Figure 3 42 Groundwater within Kakadu National Park (source: Ticknell 2008)
This is consistent with general observations about the water quality of groundwater resources of the Northern Territory including Kakadu, which are characterised by NRETA (2008) as being generally of low salinity as a result of high rainfall, high recharge rates and higher through-flow of groundwater (due to small aquifers) compared to more arid zones in the south.
3.6.6General Ecosystem/Biological Processes
Information on nutrient cycling processes within Kakadu National Park is scarce. In one study addressing nutrient cycling in the broader region, Cook (1994) investigated the effect of fires on nutrient fluxes in the tropical savannah at Kapalga, Kakadu National Park. The magnitude of nutrient fluxes due to fires was greatest in forest communities, where grassy fuel loads were high. Up to 94 percent of measured nutrients were transferred to the atmosphere during the fires. While nutrients transferred to the atmosphere as entrained ash settled within several kilometres of the fires, nutrients transferred in gaseous forms, such as nitrogen, are lost from the system. Cook (1994) noted that the losses of nitrogen greatly exceeded the inputs through rainfall and re-deposition of ash. Furthermore, nitrogen fixation was found to be of insufficient magnitude to replace the lost nitrogen, indicating that annual burning may deplete nitrogen reserves in Kakadu National Park savannas.
Similar to other areas throughout Australia’s wet-dry tropics, aquatic food webs in the Kakadu National Park region are closely linked with seasonal hydrology. Douglas et al. (2005) provide a review and conceptual model of river and wetland food webs in Australia’s wet-dry tropics, using numerous examples from Kakadu National Park. Based on this review, the aquatic food webs and associated ecosystem processes in Kakadu National Park are assumed to be underpinned by five general principles, as outlined below. Refer to Douglas et al. (2005) for further information.
Seasonal hydrology is a strong driver of ecosystem processes and food web structure. Food web structure is highly dynamic throughout the year, with the seasonal hydrological cycle (a reliable flood-pulse) driving the supply of carbon and nutrients that support food webs. The seasonal variation in water levels drives major changes in habitat availability, primary productivity and, consequently, the abundance and composition of consumer communities.
Hydrological connectivity is largely intact and supports important terrestrial-aquatic food web subsidies. The food webs of tropical rivers are characterised by very strong hydrological connections, exchanges between terrestrial and aquatic ecosystems, as well as between productive habitats like floodplains and less productive rivers habitats. Additionally, the flow regime is a key driver of exchanges between organisms and their food resources. For example, at the end of the wet season, aquatic fauna migrating from the floodplain to dry-season refuges transfer their assimilated aquatic production to these upstream rivers or billabongs. Similarly, terrestrial riparian inputs of fruit, insects, leaves and other organic debris are thought to be important contributors to the aquatic food web. The very strong aquatic linkages and floodplain connectivity are largely due to the relatively undisturbed hydrological condition of the rivers in the region (for example, lack of major waterway barriers and agriculture) and the high level of intact riparian and floodplain vegetation.
River and wetland food webs are strongly dependent on algal production. Relative to other aquatic plants and terrestrial inputs, benthic (and epiphytic) algae are typically the major source of organic carbon supporting consumers and sustaining the food webs. For instance, work in the floodplain wetlands of the East Alligator River has shown that most of the biomass carbon and nitrogen of fish and aquatic invertebrates was derived from epiphytic algae.
A few common macro-consumer species have a strong influence on benthic food webs. Tropical aquatic food webs are typically dominated by a small number of relatively large-bodied consumers, such as fish and shrimp, which have a strong influence on benthic sediments, detritus, nutrient demand, and algal and invertebrate communities. The catfish Neosiluris ater and the shrimp Macrobrachium bellatum are species that have been shown to have such an influence elsewhere in the Northern Territory, although the strengths of these top-down control effects are likely to vary in response to seasonal hydrology. The top aquatic predators within the site (for example, barramundi, saltwater and freshwater crocodiles) similarly have a strong influence on lower components of food chains.
Omnivory is widespread and food chains are short. Widespread omnivory is a characteristic of tropical fish communities, thought to be a response to the strong seasonal variability in the availability of food resources. This omnivory means that fish feed on a broad range of items, often across several trophic levels, which results in short, diffuse and highly interconnected food webs.